How does high temperature weather affect tourists’ nature landscape perception and emotions? A machine learning analysis of Wuyishan City, China

Abstract

Natural landscapes are crucial resources for enhancing visitor experiences in ecotourism destinations. Previous research indicates that high temperatures may impact tourists’ perception of landscapes and emotions. Still, the potential value of natural landscape perception in regulating tourists’ emotions under high-temperature conditions remains unclear. In this study, we employed machine learning models such as LSTM-CNN, Hrnet, and XGBoost, combined with hotspot analysis and SHAP methods, to compare and reveal the potential impacts of natural landscape elements on tourists’ emotions under different temperature conditions. The results indicate: (1) Emotion prediction and spatial analysis reveal a significant increase in the proportion of negative emotions under high-temperature conditions, reaching 30.1%, with negative emotion hotspots concentrated in the downtown area, whereas, under non-high temperature conditions, negative emotions accounted for 14.1%, with a more uniform spatial distribution. (2) Under non-high temperature conditions, the four most influential factors on tourists’ emotions were Color complexity (0.73), Visual entropy (0.71), Greenness (0.68), and Aquatic rate (0.6). In contrast, under high-temperature conditions, the most influential factors were Greenness (0.6), Openness (0.56), Visual entropy (0.55), and Color complexity (0.55). (3) Compared to non-high temperature conditions, high temperatures enhanced the positive effects of environmental perception on emotions, with Greenness (0.94), Color complexity (0.84), and Enclosure (0.71) showing stable positive impacts. Additionally, aquatic elements under high-temperature conditions had a significant emotional regulation effect (contribution of 1.05), effectively improving the overall visitor experience. This study provides a data foundation for optimizing natural landscapes in ecotourism destinations, integrating the advantages of various machine learning methods, and proposing a framework for data collection, comparison, and evaluation of natural landscape perception under different temperature conditions. It thoroughly explores the potential of natural landscapes to enhance visitor experiences under various temperature conditions and provides sustainable planning recommendations for the sustainable conservation of natural ecosystems and ecotourism.